Excel vs ERP for metal processing: where spreadsheets stop coping

Most processing plants did not choose spreadsheets. They grew into them. Stores kept a coil register, the planner built a cutting plan sheet, someone weighed scrap into a book, and the test certificates went into a box file. Over a few years the plant came to run on files that only their owners fully understand.

That is not a failure. Excel is flexible, everyone knows it, and weight arithmetic is exactly what it is good at. The question is at what point the plant needs more than a set of separate files can hold.

Where Excel works well in a processing plant

• Quotations and rate per tonne calculations, where one person builds the model.
• Slitting and cutting calculations: how many mults fit a coil width, how many blanks come out of a plate.
• One-off analysis, such as comparing yield between two suppliers' coils.
• A small plant with one machine, one planner and customers who rarely ask for certificates.
• Reports built from data that already sits in a system.

Where do spreadsheets become difficult?

Several people, several copies of stock

Stores updates the coil register, sales promises material from yesterday's copy, and the planner allocates the same coil to two orders. Shared workbooks help, but it is still easy to work from a file saved the day before.

Yield reconciled once a month

Input and output weights are written in different places: issues in the stores register, output on the production sheet, scrap in the scrap book. Bringing them together takes time, so it happens at month end, and a batch that lost too much is found long after the machine settings have changed.

Heat numbers copied by hand

Each slit coil or cut stack needs the heat number of its mother coil. In spreadsheets that means copying it onto every output line, and onto the delivery note later. One wrong digit breaks the trace, and nobody notices until a customer asks.

Offcuts and remnants

A usable plate offcut needs grade, thickness, size and heat number to be found again. Few registers keep all four, so remnants pile up and fresh plate is cut for jobs they could have served.

Production status is always behind

A spreadsheet knows only what someone typed into it. Output weights and downtime reach the office at the end of the shift, from a register or a WhatsApp photo, so the office plans on yesterday's position.

Certificates and QC records

Certificate values sit on the mill's paper, stage checks on a clipboard, and final inspection in another workbook. A customer asking for the certificate behind one delivery means opening three of them and hoping the heat numbers match.

Rework and change history

When a stack is re-levelled or re-trimmed, the record should show the original failure, the rework, the weight lost and the re-inspection. Once someone types a new stack weight over the old one, the sheet keeps no sign of the first figure or of who replaced it.

Spreadsheets and ERP compared, task by task

An ERP is an operational system, not a bigger spreadsheet

The main difference is when the data is made. A coil register is written up once the coil has already been slit. A manufacturing ERP with floor recording is filled in as the work happens: stores issues the coil to the batch, the operator records the output weights and the trim at the slitter, QC records the burr check before the coils move. Yield and stock are read from those entries, not compiled.

That is why replacing Excel with an ERP that covers only purchase, sales and the books often disappoints a processor. The files that cause the trouble are the ones about the floor, the weights and the heat numbers, and an office-only system leaves them where they were.

When should a metal processor move from Excel to ERP?

There is no tonnage at which a plant must stop using spreadsheets. The useful test is how often the files fail to give a current answer you trust. Processors who have made the move usually name some of these.

• Several people edit stock. Stores, sales and planning each keep a copy, and the same coil gets promised twice.
• More than one machine runs at once. One sheet can follow a slitter; with a cut-to-length line and a press running too, nobody has the full picture.
• Sizes and specifications change often. New widths, thicknesses and tolerances every week make the item list and conversions hard to keep right.
• Material is hard to find. The register shows a coil that is not in the bay, or an offcut that was used last month.
• Output weights arrive late. Production figures reach the office the next morning, so the plan is always a day behind.
• QC checks are on paper. Burr, flatness or hardness readings sit on clipboards, not against the batch.
• A heat number trace takes more than an hour. Answering a customer means opening the register, the production sheet and the certificate file.
• Rework cannot be seen. Nobody can say how much was re-levelled or re-trimmed last week, or what weight it cost.
• Customers ask for more. A new buyer wants certificate values with every delivery, or yield and scrap figures for a job work contract.
• WIP is unknown. Finding out what sits between operations, or at the pickling vendor, needs a walk round the bays.

None of these means Excel has failed. They mean the tonnage, the machines and the customers have outgrown a register, a cutting sheet and a scrap book kept apart. When two or three show up together, the registers are usually eating more hours than a system would. The guide to implementing ERP in a metal processing plant covers what the move involves.

Keeping what works

Excel stays in the plant after an ERP arrives. Rate calculations, slitting arithmetic and one-off analysis stay in spreadsheets, using exports where needed. What moves is the operational record: stock by lot and heat number, batches with their weights, scrap by reason, checks and dispatch. In FleekERP that record sits on the work order, with inventory and lot traceability and yield and scrap reports in manufacturing intelligence, while the books stay in Tally or Zoho.

For how heat numbers, batch yield and scrap by reason run on one record, see FleekERP for metal processors.